1. Total heat loss was measured by 24 h direct calorimetry in five obese and five lean women who were maintained throughout the study on a diet supplying 3·3 MJ/d. Each subject was measured five times to assess the effect of temperature, exercise and food on energy expenditure. Within each weight group a Latin-square design was used to balance sequence effects on the thermogenic responses to temperature, exercise and food.
2. Compared with the control day, on which no thermogenic stimulus was given, the increase in 24 h heat production by the lean and obese women caused by 30 min exercise on a bicycle ergometer against a load of 20 N was 10·1 and 10·3 W for obese and lean groups respectively. There was no evidence in either group of a measurable long-term increase in metabolism which would increase the energy cost of the exercise above that predicted from indirect calorimetry during the exercise.
3. The increase in heat production associated with ingesting an extra 4·4 MJ (obese group) or 4·0 MJ (lean group) was 3·4 and 3·0 W respectively. This response was similar to that predicted from indirect calorimetry for a few hours after the meal.
4. The obese and lean groups differed in metabolic response to calorimetry at the upper or lower limits of the thermal comfort zone, which was determined individually for each subject. The difference from control values in the obese group was an increase of 3·8 W on the ‘warm’ run, and a decrease of 2·0 W on the ‘cool’ run. Among subjects the change was an increase of 0·4 W on the ‘warm’ run, and an increase of 4·8 W on the ‘cool’ run. The differences between the groups did not achieve statistical significance. The lower and upper temperature limits were similar in the two groups: 23·2–26·4° for the obese group, and 23·3–26·2° for the lean group.
5. The most striking difference between lean and obese subjects in the present study was the much higher resting metabolic rate, and total energy expenditure, of the obese group. During the control run the obese group had a mean energy expenditure of 96·1 W, compared with 61·7 W in the lean group. There was no overlap: the lowest energy expenditure for an obese subject was 81·4 W and the highest for a lean subject was 76·1 W. In comparison to this large difference in baseline the magnitude of the thermogenic responses was small.
6. Under the conditions of this study there was nothing to support the view that a failure of thermogenic response is an important factor in the causation of human obesity. To support that view it would be necessary to show differences in thermogenesis in lean and obese subjects which were at least an order of magnitude greater than those which we have observed.